Journal of Natural Pmducts Vol. 55. No. 8.@ 113G-1141. Augur 1992
1136
(+)-4P-"YDROXYHERNANDULCIN, A N E W SWEET SESQUITERPENE FROM THE LEAVES AND FLOWERS OF LIPPIA DULCIS' NORITO KANEDA, IK-%O LEE,MAHABIRP. GUFTA,2 D. DOELSOEJARTO,and A. DOUGLAS KINGHORN+ Program fw Collahrative Research in the Pharmaceutical Sciences, Department of Medicinal C h i s t r y and Pharnarognosy, College of Pharmacy, Univrrsity of Illinois at Chicago, Chicago, Illinois 60612 AsSTUa.-From the leaves and flowers of Lippia &Iris collected in Panama, a new sweet sesquiterpene identified as (+)-4p-hydroxyhernandulcin 121was isolated, accompanied by (+)hernandulcin [l],(-)-epihernandulcin 131 (a novel natural product), and 6-methyl-5-hepten2-one [4].Acteoside (verbascoside)IS],a known bitter phenylpropnoid glycoside, was isolated from the flowers of L. dulcis. The structure of (+)-4p-hydroxyhernanddcin was established by interpretation of its spectral data.
(+)-Hernandulcin E l ] was first isolated as a sweet bisabolane sesquiterpene constituent of Lippta dulcis Trev. (Verbenaceae) leaves and flowers collected in Mexico (2). It was rated by a taste panel as being 1000 times as sweet as sucrose on a molar basis (2). The racemic form of this compound was synthesized by directed aldol-condensation starting from the two commercially available ketones, 6-methyl-5-hepten-2-one and 3methyl-2-cyclohexen- 1-one. (f)-Hernandulcin was neither acutely toxic for mice at single doses of up to 2 glkg body wt, nor positive in a forward mutation assay, when evaluated both in the presence and absence of a metabolic activating system (2,3). Attempts to synthesize derivatives of hernandulcin El]with improved hedonic parameters resulted in the discovery of no further sweet-tasting analogues. However, as a result of
2
4
3
1 R=H R=OH
6"'
5"'
5 'Paper No. 26 in the series, "Potential Sweetening Agents of Plant Origin." For part 25, see Fullas e~ al. (1). 2Address: FLORPAN, Unidad de Investigaciones Farmacognosticas, Universidad de PanamL, Panama, Republic of Panama.
August 19921
Kaneda et af.: Sweet Sesquiterpene
1137
such investigations, it was concluded that the C-1 keto group, the C-1’ tertiary hydroxyl group, and the double bond between C-4‘ and C-5’ must be present in order for compound 1 to exhibit sweetness (4). Both (+)- and (2)-hernandulcin [l]have been synthesized by other groups (5-10). Thus, Mori and Kat0 (5,6) have synthesized (+)-hernandulcin El} from (R)-(+)limonene, and thereby established that the absolute configuration of naturally occurring hernandulcin El] is bs,1’s.Furthermore, they showed by synthesizing the other three stereoisomers 1’R; 6R, 1’R; and 6R, 1’s)of this compound from appropriate isomers of limonene that only the bs,1’s enantiomer is sweet (6). Racemic hernandulcin El] has been synthesized variously from a cyclohexadiene derivative using boron and silicon enolates (7), by an intramolecular nitrile oxide cycloaddition route from (22, bE)-farnesal(8 ,9), and from an E-dienyl carbonate by titanium chloride catalysis (10). The naturally occurring form of this sweet sesquiterpene, (+)-hernandulcin E l ] , has been produced from both hairy root cultures (1 1)and shoot cultures (12) of L. dukcis. In the present communication, we report the isolation and characterization of a second sweet bisabolane sesquiterpene, the novel (+)-4P-hydroxyhernandulcin 121, along with several analogues, including (+)-hernandulcin El] in high yield, from L. dufcis leaves and flowers collected in Panama. A sample ofL. dulcis {which is listed in the Ffma of Panama under the name Pbyka scaberrima (A.L. Juss.) Moldenke} (13), was obtained from a medicinal plants’ market in Valle de A n t h , CoclC, where it was on sale for the treatment of respiratory ailments.
(a,
RESULTS AND DISCUSSION The molecular formula of compound 2 was determined as C15H2403 (m/z 252.17 15) from its hrms data, indicating that i t was a structural analogue of (+)-hernandulcin 111. On comparison of its [ a ] ~ I3C-nmr , (6 67.5), and eims spectra ([MI+ 252 ) with those of 1 ([MI+ 236), 2 was considered to be a hydroxylated derivative. Initial spectral assignments of compound 2 were made after ‘H-, 13C-nmr, ‘H- ‘H COSY, and *H-13C HETCOR nmr experiments were performed. The position of the additional hydroxyl group in 2 was deduced as being attached to C-4, on the basis of resonance enhancements at chemical shift values corresponding to C-2, C-3, C-6, and C-7, which were observed upon irradiation ( v c H = 6 H t ) of H-4 (6 4.26) in a selective INEPT nmr experiment (14). The assignment of the hydroxyl group at C-4 position was confirmed by proton homonuclear decoupling nmr experiments. Thus, irradiation ofH-5axat6 1.91 transformedH-4(64.26, ddd,J=3.1, 3.1, 6.3Hz)toanarrower multiplet peak similar to a broad singlet, and caused H-6 (6 2.90, dd, J = 12.9, 4.6 Hz) to collapse to a doublet with a smaller] value (4.6 Hz). When H-5eq (6 2.11) was irradiated, both the H-4 and H-6 peaks were simplified to a narrower multiplet and a doublet with a IargerJ value (12.9 Hz), respectively. An analysis of the individual spin system constituted by H-4, H-5, and H-6 led to the conclusion that the stereochemistry of hydroxyl group at the C-4 position is P (axial), a summary of which is shown in Figure 1. Thus, based on the observed coupling constants between H-4 and H-5 and between H-5 and H-6, together with a consideration of the absolute stereochemistry of 1, in which H-6 is in the P configuration (5,6), it may be suggested that the actual splitting pattern (pseudo-quintet) of H-4a is caused by the spin-spin interactions between H 4 a ( e q ) and H-5P(eq) ( J = 3. l Hz), H-4a(eq) and H-Sa(ax) ( J = 3. l Hz), and H 4 a ( e q ) and 4P-OH(ax) ( J = 6.3 Hz). Likewise, the spin system of H-bP(ax) was influenced by H-5P(eq) ( J = 4 . 6 Hz), and by H-Sa(ax) ( J = 12.9 Hz), to give a well resolved doublet of doublets signal. Further supportive evidence for the stereochemistry of 4P-hydroxyhernandulcin [2] was obtained by comparing the observed ‘H- ‘H
Journal of Natural Products
1138
(Vol. 55, No. 8
SI, 1.26 (ddd. .I = 3.1. 3.1, 6.3 Hz)
2
H ax
FIGURE 1.
(9")
H a
A partial perspective view and Newman projection models of compound 2.
coupling constants between H-4 and H-5 and between H-5 and H-6, with those predicted by molecular mechanics calculations, in whichJ values were given as 2.6 Hz forJ~4a-H5a(~), 3.6 Hz fofJH4a-H5p(eq), 3.5 Hz for]H6p-H5p(eq)r and 12.3 Hz for JH6p-H5a(ax). Therefore, the chemical structure of this new sesquiterpene was elucidated as (+)-4P-hydroxyhernandulcin (27. The sweetness potency of this compound relative to sucrose was not determined because of the small amount isolated. 4P-Hydroxyhernandulcin (21 is significant in being only the second naturally occurring or synthetic hernandulcin analogue reported to have a sweet taste. (+)-Hernandulcin 111 was obtained in a much higher overall yield in this Panamanian sample of L . dulcis (0.154% w/w) than in the sample collected in Mexico (0.004% w/w) (15). It is interesting to note that there was no evidence of camphor in the Panamanian specimen; camphor, however, was the major volatile oil constituent of L. dulcis collected in Mexico (15). More precise 'H-nmr coupling constant assignments were made in the present investigation for 1 than those published previously (2). Compound 3 exhibited levo-rotation at 589 nm, but it resembled (+)-hernandulcin in its other spectral data (uv, ir, 1H and 13Cnmr) and exhibited the same base (m/z 237) and fragment peaks ( d z 2 19, 109) in the eims spectrum, suggesting the possibility of its being a diastereomeric counterpart because of its ease of separation by tlc. It was established as (-)-epihernandulcin 131 (a new natural product) by comparison of spectral data with the (*)-,(+)-, and (-)-forms of this compound, which have been obtained previously by chemical synthesis (2,3,6). (-)-Epihernandulcin (31, the absolute configuration ofwhich was assigned as 6R, l'S, has been determined as being devoid of sweetness in a previous study (6). Compound 4 was identified as a known compound, 6methyl-5-hepten-2-one, from its 'H- and 13C-nmrdata. Although hernandulcin 111 is known to degrade to 4 by thermal dissociation at 140" (3), there was no evidence of its formation when an aqueous solution of 1 was stored at room temperature for two weeks at neutral pH. Accordingly, it may be inferred that 4 is a natural product. The methanol extract of L . dulcis flowers was examined to see if any glycosides of hernandulcin (11were present. Although no such compounds were detected, the major constituent of the MeOH-soluble extract was the known phenylpropanoid glycoside, acteoside (verbascoside)(51, which was identified by comparison with published physical and spectroscopic data (16-19). This compound was initially characterized by Birkofer
August 19921
Kaneda et al.: Sweet Sesquiterpene
1139
et a / . (20) and was obtained as a bitter-tasting principle of Conandronramidioides ( G e s -
neriaceae) (2 1). This substance has been recently isolated from Lantana camara, another plant in the family Verbenaceae (22), and exhibits significant cytotoxicity for P-388 cells (23) and antimicrobial, immunosuppressive, and protein kinase C inhibitory effects (22).
EXPERIMENTAL GENERAL EXPERIMENTAL PROCEDURES.*tiCd rotations were measured with a Perkin-Elmer Model 24 1 polarimeter. Uv spectra were obtained on a Beckman DU-7 spectrometer. Ir spectra were taken with a Midac Collegian FT-IR spectrophotometer. ‘H-nmr and I3C-nmr spectra were measured with TMS as internal standard, employing a Varian XL-300 instrument operating at 300 MHz and 75.6 MHz, respectively. ‘H-’H COSY and ‘H-13C HETCOR nmr experiments were also performed on a Varian XL300, using standard Varian pulse sequences. Selective INEPT nmr experiments were conducted on a Nicolet NT-360 spectrometer operating at 90.8 MHz. Eims, cims, and hreims were obtained using a Finnigan MAT 90 instrument. Conformational analysis in order to compute nmr coupling constants was performed by optimization of structures using molecular mechanics energy calculations, with the use of MacroModel software (24), run on a Digital Equipment Corporation VAX 111785 computer interfaced to a Tektronix 4 107 graphics display terminal. PLANTMATERIAL.-The leaves and flowers of L. dukzs were purchased in May 1991, from a marketplace in Valle de A n t h , CoclC, Panama. The identity of the sample was confirmed by a taxonomist (D.D.S.). A voucher specimen (Florpan 743) representing this collection has been deposited at the Herbarium of the Field Museum of Natural History, Chicago. EXTRACTIONAND ISOLATION.-The airdried and powdered aerial parts (leaves and flowers) (1.1 kg) of L. dulcis were exhaustively extracted with petroleum ether (6 X 2 liters). The petroleum ether layer was combined with that obtained from a similar extraction of L. &Iris flowers (70 g). after confirming the similarity of their tlc profiles. This combined petroleum ether layer was concentrated to dryness, and the residue (25 g) was chromatographed over a Si gel column (500 g) by elution, in turn, with toluene1hexane. toluene, toluene/Me,CO, and toluene/Me,CO/MeOH, with (+)-hernandulcin [l]to guide the fractionation. Additional Si gel cc with toluene-EtOAc (15: l) as eluent was carried out on the second fraction (5.6 g) from the original column, obtained from elution with toluene, to afford 6-methyl-5-hepten-2-one 141 (0.5 g, 0.043% w/w)and(+)-hernandulcin[l](1.8g, 0.154% w1w). Fromthethirdfraction(l.8g)of the original cc, eluted with toluene/Me,CO, (-)-epihernandulcin [3] (0.47 g, 0.040%wlw) was purified by further cc over Si gel with cyclohexane-EtOAc (5: 1). Likewise, from the fourth fraction (3.7 g) of the original cc eluted with toluene/Me,CO/MeOH, 2 (9.5 mg, 0.0008% wlw) was obtained by subsequent cc using CHCI3-Me,CO (15: 1) and toluene-EtOAc ( 2 : l ) as solvents. This compound exhibited a similar orange-brown color in visible light to 1 when visualized with 10% H2S04(heated at 1lo”, 10 min). The air-dried flowers (70 g) of L. &Iris were exhaustively extracted with MeOH (3 X 1 liter) at room temperature, and MeOH was removed under vacuum. This extract was taken up in MeOH (50 ml), and the resultant suspension in H,O (50 ml) was washed consecutively with petroleum ether (3 X 100 ml) and Et,O (3 X 100 ml). The aqueous MeOH layer was then treated with n-BuOH (2 X 100 ml). The n-BuOH layer was concentrated to dryness, and the residue (3.5 g) was subjected to chromatography over a Si gel (200 g) column with CHC1,-MeOH (8: 1) as eluent, to afford acteoside (verbasccside) [5] (1.6 g, 2.3% w1w).
+
(+)-Hernandulcin [l].-Colorless oil: [a]”D 105.1” (c= 3.3, CHCI,); ‘H nmr (CDCI,) 6 1.17 (3H, s,H-8’), 1.47 (2H, dd,]=8.3, 8.3 Hz, H-2’), 1.62(3H, s, H-7‘). 1.70(1H, m, H-Sax), 1.68 (3H, s, H-6’), 1.98 (3H, s, H-7), 2.03 ( l H , m, H-Seq), 2.05 ( l H , m, H-3’), 2.13 ( l H , m, H-3’), 2.30 ( l H , m,H-4), 2.35(1H,m,H-4), 2.42(1H,dd,]= 14.0,4.6Hz,H-6),5.12(1H,tt,]=7.2, 1.3Hz, H-4’), 5.30(1H, s, l’-OH), 5.88 ( l H , s, H-2); eimsm/z(rel. int.)[M+ 1]+ 237 (1001, 219 (GI),103 (78). For uv, ir and I3C-nmr spectral data of [l],see Compadre and co-workers (2,3).
+
(+)-4P-Hydrmyhernandulrzn [Z).--Colorless oil: [ a ] 2 3 ~ 103.0° (c= 0.77, CHCI,); uv A max (EtOH) (log e) 232 (4.29) nm;ir u max (film) 3405 (br), 2970, 2920, 2860, 1650, 1440, 1380, 1220 cm-’; ‘Hnmr(CDC1,)6 l . l 6 ( 3 H , s, H-8’), 1.48(2H, dd,J=8.4,8.4Hz, H-2’), 1.63(3H,s, H-7’), 1.68(3H,s,H-6’), 1.91(1H,ddd,]= 13.4, 13.4, 3.7Hz,H-5ax),2.07(3H,s,H-7),2.11(1H,m,HSeq), 2.14 ( l H , m, H-3’), 2.18 ( l H , m, H-3’), 2.33 ( l H , d,]=5.6 Hz, 4-0H), 2.90 ( l H , dd, ] = 1 2 . 9 , 4 . 6 H ~ , H - 6 ) , 4 . 2 6 ( l H , d d d , / = 3 . 1 , 3 . 1 , 6 . 3 H z , H - 4 ) , 4 . 9 8 ( 1 H , s , l’-OH).5.11 ( l H , tt,] = 6.6, 1.3 Hz, H-4’), 5.84 ( l H , s, H-2); I3C nmr (CDCI,) 6 18.0 (q, C-7’), 2 1.9 (t, C-3‘), 22.0 (q, C-7), 24.2 (4, C-8’), 26.1 (4,C-6’). 33.6 (t, C-5), 40.5 (t, C-2’), 46.9 (d, C-6). 67.5 (d, C-4), 74.3 (s, C-If), 124.7 (d, C-4’), 128.5 (d, C-2), 132.0 (s, C-5’), 159.8 (s, C-3), 203.9 (s, C-1); eims m h
1140
Journal of Natural Products
[Vol. 5 5 , No. 8
(rel. int.) [MI+ 252 (6), 234 (16). 169 (15). 151 (229, 125 (58), 109 (loo), 82 (25), 69 (54), 55 (31); hreims m/z [MI+ 252.1715 (calcd for C15H2403 252.1725).
( - ) - E p i h n b l c i n [3].-€olorless oil; [a]*'D - 180.6" (c= 1.03, CHCI,) [lit. (6) [aI2'D - 133" 0.108, EtOH)]; uv A max (EtOH) (log E) 237 (4.13) nm; ir v max (film) 3440, 2970, 2915, 2860, 1650, 1430, 1380, 1215 cm-'; 'H nmr (CDCI,) 6 1.19 (3H, s, H-8'), 1.40 ( I H , ddd, J = 12.7, 12.7, 5.0 Hz, H-2'), 1.54 ( I H , dd,]= 12.1, 4.9 Hz, H-2'), 1.60 (3H, S, H-7'), 1.66(3H, S, HL6 1.75 ( I H , dddd,]= 12.8, 12.8, 12.8,5.6Hz, H-Sax), 1.89-1.11(2H,m,H-5eqandH-3', 1.97(3H,s, H7), 2.18(1H,m,H-3'), 2.24-2.40(2H,m, H4),2.35(1H,dd,]=13.7,4.8Hz,H-6), 5.08(1H,s, I'-OH), 5.09(1H,m,H-4'), 5.84(1H,s,H-2); L3Cnmr(CDC13)617.6(q,C-7'),22.l(t,C-3'), 24.1 (q, C-7), 25.0(t, C-5), 25.4(q,C-8'), 25.7(q,C-6'), 3 1 . 5 ( t 9 C 4 ) ,37.1(t, C-2'), 55.3(d,C-6), 7 4 . 3 ( ~ , C-l'), 124.8 (d, C-4'), 127.4 (d, C-2). 131.1 (s, C-5'), 163.6(s, C-3), 203.4(s, C-l);eimsm/z(rel. int.) [M I]+ 237 (loo), 2 19 (88), 109 (70). (c=
+
6-Methyl-5-hcptm-2-one[4].-Colorless oil: ir Y max (film) 2970, 2920, 2860, 1720, 1440, 1360, 1160 cm-'; 'H nmr (CDCI,) 6 1.62 (3H, s, H-8), 1.67 (3H, S, H-7), 2.13 (3H, S, H-l), 2.25 (2H, dd, ]= 14.5, 7.5 Hz, H-4), 2.46 (ZH, t, ]=7.4 Hz, H-3). 5.07 ( I H , tt,]= 7.1, 1.4 Hz, H-5); 13C nmr (CDCI,) 6 17.3 (9, C-8), 22.3 (t, C-4), 25.3 (q, C-7), 29.5 (q, C- I), 43.4 (t, C-3), 122.5 (d, C-5), 132.3 (s, C-6), 208.2 (s, C-2); cims m/z (rel. int.) [M I]+ 127 (93), 125 (5 l), 119 (61), 117 (57). 109 (100).
+
Acteoside (=verbascoside) [5].-Amorphous powder: [a12'D -82.8" (c= 0.39, MeOH) [lit. (17) (c= 4.0, MeOH); lit. (16) [ a I 2 ' ~-66.5' ( c = 1.0, MeOH)]; fabms m/z [M + I]+ 625. For uv and ir spectral data of 5 , see Kitagawa et al. (16) and Lahloub et al. (18); for 'H-, and l3C-nmr spectral data, see Jia et al. (17), Lahloub et al. (18), and Andary et al. (19). [ a ] 1 5-70.2' ~
ACKNOWLEDGMENTS This investigation was funded by grant R01-DE-08937 from the National Institute of Dental Research, NIH, Bethesda, Maryland. Professor Mireya Correa, of the University of Panama, is thanked for providing valuable input on the taxonomic aspects of the material used in this investigation. Thanks are also due to the Regional Program for Scientific and Technological Development of the Organization of American States and the Commission of European Communities for supporting FLORPAN in Panama. We acknowledge the Nuclear Magnetic Resonance Laboratory of the Research Resources Center, University of Illinois at Chicago, for expert assistance and for provision of spectroscopic equipment used in this study. We thank Dr. M. E. Johnson for useful suggestions concerning the use ofMacroMode1software, and Mr.R.B. Dvorak, ofthe Department ofMedicina1Chemistry and Pharmacognosy, University ofIllinois at Chicago, for recording the mass spectral data. LITERATURE CITED 1.
2. 3. 4. 5.
6. 7.
8. 9. 10.
11. 12. 13. 14. 15. 16. 17. 18.
19.
F. Fullas, R.A. Hussain, E. Bordas, J.M. Pezzuto, D.D. Soejarto, and A.D. Kinghorn, Tetrahedron, 4 7 , 8515 (1991). C.M. Compadre, J.M. Pezzuto, A.D. Kinghorn, and S.K. Kamath, Science, 227, 4 17 (1985). C.M. Compadre, R.A. Hussain, R.L. Lopez de Compadre, J.M. Pezzuto, and A.D. Kinghorn,]. Agric. FoodChem., 35, 273 (1987). C.M. Compadre, R.A. Hussain, R.L. Lopez de Compadre, J.M. Pezzuto, and A.D. Kinghorn, Experientia, 44,447 (1988). K. Mori and M. Kato, Tetrahedron Lett., 27, 981 (1986). K. Mori and M. Kato, Tetrahedron, 42, 5895 (1986). Y.N. Bubnov and M.E. Gurskii, Izv. Akad. Nauk SSSR, Ser. Khim., 1448 (1986); Chem. Abstr., 1 0 7 , 4 0 1 0 2 ~(1987). G.-C. Zheng and H. Kakisawa, Bull. Chem. Sor. Jpn., 6 2 , 602 (1989). G.-C. Zheng and H. Kakisawa, Chin. Sci. Bull., 35, 1406 (1990). P.F. De Cusati and R.A. Olofson, Tetrahedron Lett., 31, 1409 (1990). M. Sauerwein, T. Yamazaki, and K. Shimomura, Plant CellRep., 9 , 579 (1991). M. Sauerwein, H.E. Flores, T. Yamazaki, and K. Shimomura, Plant Cell Rep., 9 , 663 (1991). H.N. Moldenke, Ann. Missouri Bot. Gard., 60, 60 (1973). A. Bax, J . M a p . Reson., 57, 314 (1984). 1 5 , 89 (1986). C.M. Compadre, E.F. Robbins, and A.D. Kinghorn,]. Ethnopha-ol., S. Kitagawa, H. Tsukamoto, S. Hisada, and S. Nishibe, Chon. Pbann. Bull., 32, 1209 (1984). Z. Jia, Z. Liu, and C. Wang, Pbytorhistv, 30,3745 (1991). M.F. Lahloub, A.M. Zaghloul, S.A. El-Khayaat, M.S. Afifi, and 0. Sticher, Planta Mu'. , 5 7 , 4 8 1 (1991). C. Andary, R. Wylde, C. Laffite, G. Privat, and F. Winternitz, Phytorhmirrry, 21, 112 (1982).
August 1992)
Kaneda et al. : Sweet Sesquiterpene
1141
20. L. Birkofer, C. Kaiser, and U. Thomas, Z . Natwfwscb., Tn’l B , 23, 105 1(1968). 2 1. G. No& and I. Nishioka, Phytorhrmistry, 16, 1265 (1977). 22. J.M. Herbert, J.P. Maffrand, K. Taoubi, J.M. Augereau, I. Fouraste, and J. Gleye, J . Nat. Prod., 54, 1595 (1991). 23. G.R. Pettit, A. Numata, T. Takemura, R.H. Ode, A.S. Narula, J.M. Schmidt, G.M. Cragg, and C.P. Pase, J . Nat. Prod., 53,456(1990). 24. F. Mohamadi, N.G.J. Richards, W.C. Guida, R. Liskamp, M.Lipton, C. Caufield, G. Chang, T. Hendrickson, and W.C. Still, J. C m p t . C h . ,11,440(1990). R e c e z d 24 Fcbnrrrry 1992
ERRATUM For the paper by Dillman and Cardellina entitled “Aromatic Secondary Metabolites from the Sponge Tedunia ignis,”J . Nat. Prod., 54,1056 ( l w l ) , structure 3 should be a p-carboline, not a carbazole. The corrected structure drawing is as follows:
